FR2985120A1 - METHODS FOR TRANSMITTING AND RECEIVING DATA SYMBOLS - Google Patents

Abstract

According to the invention, the transmission method by a transmitting device comprising M transmitting antennas to a receiving device comprising N + C receiving antennas, M, N and C being integers greater than or equal to 1, comprises: - a step (E10) of spatial frequency multiplexing of a block of N data symbols D, D, ..., D resulting in a block of N useful spatial symbols S, S, ..., S, this step using an inverse discrete Fourier transform (IDFT); a step (E20) of adding C space redundancy symbols to the block of N useful spatial symbols, resulting in a block of N + C spatial symbols S, S, S; and a step (E30) of precoding the N + C spatial symbol block S, S, S by means of a focusing matrix of dimensions Mx (N + C), providing M precoded spatial symbols X, X, ..., X, each precoded spatial symbol being transmitted on a separate transmitting antenna.

Description

BACKGROUND OF THE INVENTION The invention relates to the general field of telecommunications, and in particular to the field of digital communications.

It relates more particularly to the transmission (transmission and reception) of signals formed of data symbols by a digital communications system equipped with multiple receiving antennas, also known as the SIMO system (Single Input Multiple Output) when the system includes only one transmit antenna or Multiple Input Multiple Output (MIMO) system when the system is equipped with multiple transmit antennas. The invention thus applies in a privileged manner to wireless communications, for example on a fourth or future generation network (eg LTE networks, Long Term Evolution) as defined in the 3GPP standard (Third Partnership Project), or on optical transmission networks. It can be applied both downlink (from a base station to a terminal) and uplink (from a terminal to a base station). Furthermore, the invention applies in a privileged manner to a flat frequency propagation channel, that is to say a channel without echoes (or "single tap" in English), also known as a channel. "Flat fading" in English. In a known manner, the multi-antenna systems make it possible to achieve very high transmission rates, the capacity of the MIMO channels increasing proportionally with the minimum number of transmit antennas and receiving antennas. Among these multi-antenna systems, the systems using a large number of antennas (or "large scale multiple antenna networks" in English), typically of the order of one or several hundred antennas, collocated at the level of the same site or distributed geographically, offer many advantages such as higher speeds, increased reliability of transmission, energy savings, etc. Various data symbol transmission techniques have been proposed for MIMO systems, in the case of a single-carrier system, with a flat fading channel. In the particular case of a "flat fading" channel, the propagation channel between any antenna of the transmitter and any antenna of the receiver can be simply modeled by a complex gain. As a result, the propagation channel between a transmitter provided with a plurality of transmitting antennas and a receiver equipped with several receiving antennas is written in the form of a complex matrix, called matrix of the MIMO propagation channel, of which each line corresponds to a reception antenna, and each column to a transmitting antenna.

Among these techniques, some rely on knowledge of the matrix of the MIMO transmission channel in transmission. This knowledge of the channel makes it possible to calculate a focusing matrix Q (or "beamforming" matrix), which is applied to the data symbols before they are transmitted on the transmission antennas. This precoding matrix allows a focusing of each data symbol on a particular receiving antenna to facilitate the decoding in reception of the received data symbols. Different forms of focusing precodings can be envisaged in transmission. Thus, for example, one can consider a precoding focus by time reversal. In the case of a "flat fading" channel, the time reversal consists in using as the focusing matrix the conjugate transpose of the MIMO propagation channel matrix. In the remainder of the description, the concept of the transmission or propagation channel between transmitting antennas and receiving antennas includes not only the effects of the medium on which the digital signal propagates between the transmitting antennas. and reception antennas (eg radio or wired channels) but also the effects induced by the transmit and receive antennas on the digital signal. The implementation of such a transmission pre-encoding matrix makes it possible to use in reception a simple equalization scheme on each antenna. Indeed, the matrix of the equivalent channel resulting from the product of the precoding matrix and the matrix of the propagation channel MIMO, is a quasi-diagonal matrix. Therefore, no complex matrix inversion in reception is needed to decode the received signal on the receiving antennas. On the other hand, obtaining the precoding matrix does not require complex matrix inversion in transmission (conjugated transpose of the matrix of the MIMO channel only).

Such a transmission scheme is therefore particularly well suited for MIMO systems based on large antenna arrays. However, it can be observed that with this transmission scheme, interference between the data symbols remains in reception (these interferences correspond to the non-diagonal and non-zero terms of the matrix of the equivalent channel), which has an impact in terms of performance. In other words, with this transmission scheme, each symbol is focused on a separate target receiving antenna creating a focal spot that is centered on that target antenna, but interferes with neighboring party antennas. Other MIMO transmission systems have been proposed to reduce interference between residual symbols. These systems are for example based on the use of ZDF or ZF for Zero Forcing precoders: these precoders make sure to cancel the non-diagonal terms of the matrix of the equivalent channel, and that a focal spot centered on a target receiving antenna has zero interference with neighboring antennas; or - minimizing the mean square error (MMSE). However, these systems require complex matrix operations in reception, notably involving matrix inversions. On the one hand, the feasibility of such inversions is not guaranteed. On the other hand, these matrix inversions are not practical in practice for a large number of antennas. There is therefore a need for a digital transmission scheme that can be used in the context of multi-antenna networks having a large number of antennas and not having the disadvantages of the state of the art. OBJECT AND SUMMARY OF THE INVENTION The invention responds in particular to this need by proposing a transmission method and a method for receiving signals intended to be implemented in a system comprising M transmit antennas and N + C antennas. reception, M, N and C being integer parameters greater than or equal to 1, for the transmission of N data symbols. More specifically, the invention proposes a method of transmitting data symbols by a transmitting device comprising a number M of transmitting antennas, to a receiving device comprising a number N + C of receiving antennas, M, N and C being integer parameters greater than or equal to 1, this transmission method comprising: a step of multiplexing in spatial frequency of a block of N data symbols D1, D2,. a block of N useful spatial symbols S ', 1.5u, 2, .-., S ,,, N, this step using an inverse discrete Fourier transform; a step of adding C spatial redundancy symbols to the block of N useful spatial symbols, resulting in a block of N + C spatial symbols S1, S2, SN + C; and a step of precoding the block of N + C spatial symbols S1, S2, SN + c by means of a focusing matrix of dimensions Mx (N + C), said precoding step providing M precoded spatial symbols X1, X2, ..., XM, each precoded spatial symbol being transmitted on a transmitting antenna distinct from the transmitting device.

The invention also relates to a reception method, by a receiving device comprising a number N + C of receiving antennas, N + C spatial symbols 1. YY - r - 2r - - -, YN + C these space symbols received resulting from the propagation via a propagation channel of M spatial symbols X1, X2,..., XM transmitted by a transmitting device via a number M of transmitting antennas, these precoded spatial symbols X1, X2,. XM being derived from a precoding, using a focusing matrix, a block of N + C spatial symbols S1, 52, SN + c, these spatial symbols S1, 52, ..., SN c themselves being derived from a spatial frequency multiplexing of data symbols D1, D2,..., DN using an inverse discrete Fourier transform and an addition of C spatial redundancy symbols, M, N and C being integer parameters greater than or equal to 1. According to the invention, the reception method comprises: a step of extracting N useful spatial symbols received S'u, 2, -., S'U, N among the N + C space symbols received Y1, Y2, -., YN + C; a step of spatial frequency demultiplexing of the N useful spatial symbols received ..., S'U, N, resulting in a block of. N demultiplexed data symbols D'2, ..., D'N, this step using a discrete Fourier transform; and a step of equalizing the block of N demultiplexed data symbols D'2, D'N.

Thus the invention proposes to judiciously combine different processing on transmission to allow, at the receiver, to reduce the effect of interference created by a focal spot centered on a so-called target receiving antenna (spot which reflects the level of the signal intended for this target antenna received on the receiving antennas of the receiving device), on the neighboring antennas. These processes include a spatial frequency multiplexing of N data symbols in a block of N useful spatial symbols, an addition of C spatial redundancy symbols to useful spatial symbols and a precoding allowing a focus of the symbols transmitted on the N + C antennas reception. More specifically, the invention makes it possible to exploit a phenomenon of "hardening" of the propagation channel (also known as "channel hardening" in English) around the reception antennas induced by the transmission scheme of the invention, and which manifests itself in particular by an invariance of certain statistics of the channel at the receiving antennas. This hardening of the propagation channel results in a focal spot at each substantially identical reception antenna regardless of the position of the receiving antenna.

This focal spot for a target receiving antenna spreads over several receiving antennas around the target receiving antenna, and interferes with these antennas. However, the addition of C spatial redundancy symbols in transmission to useful spatial symbols resulting from a spatial frequency multiplexing of data symbols according to the invention, combined with the use of N + C receiving antennas at the of the receiver device, advantageously makes it possible to reduce the effect of these interferences at the level of the receiver, after extraction of the useful spatial symbols received. In other words, the invention uses C receiving antennas to absorb the interference, while N receiving antennas are dedicated to the reception and decoding of the useful signals transmitted by the transmitting device. The step of extracting the N symbols S'U, N makes it possible to eliminate the interferences. The step of adding C redundancy symbols during the transmission process of the invention may comprise in particular: the insertion of a suffix of LT spatial redundancy symbols following the block of N useful spatial symbols ; and / or - the insertion of a prefix of KT spatial redundancy symbols at the head of the block Sto, Su, 2, -, S ,, N of N useful spatial symbols; LT and KT designating two positive or zero integer parameters such that C = KT + LT. In a particular embodiment, KT-C, LT = 0 or KT = 0 and LT = C.

Alternatively, other configurations distributing spatial redundancy symbols at the beginning and at the end of the Su, I, Su, 2, -, Su, N block of useful spatial symbols can of course also be considered. The added redundancy spatial symbols do not contain useful information as such. Preferably, a prefix and / or a cyclic suffix (s) is added to the block of N useful spatial symbols: thus, the added prefix takes up the last KT symbols of the block of N useful spatial symbols Sua, Su, 2, -, Su, N while the added suffix takes up the first LT symbols of the block of N useful spatial symbols SU, 1, Su, 2, -, S ,,, N. Moreover, a judicious choice of the focusing matrix used in transmission for the precoding of spatial symbols makes it possible, on the one hand, to limit the complexity of precoding on transmission, and on the other hand to limit the complexity of the equalization performed in reception. Thus, for example, the focusing matrix may be a time reversal matrix obtained by the conjugate transpose of an estimate of the matrix of the MIMO propagation channel between the transmitting antennas M of the transmitter device and the N + C antennas of reception of the receiving device. The use of such a precoding matrix, combined with the insertion of C redundancy symbols in the form of a prefix and / or a cyclic suffix (s), advantageously makes it possible to transform the channel between the N + C spatial symbols S1, S2, SN-Fc and the N + C spatial symbols received, in an equivalent channel, between the N useful spatial symbols SU, 1, SU, z, ..., SU, N and the N symbols Spatial spatial data received S ',, 1, S' ,, 2, S'U, N, modeled by a circulating matrix. Such a matrix is, in known manner, diagonalizable in a Fourier base. This has the consequence that after demultiplexing the useful spatial symbols received S'10,..., S'U, N, a very simple equalization scheme that does not require a matrix inversion can be implemented. Other focusing matrices can be used to implement the invention. Thus, in another variant, the focusing matrix may be a matrix of ZF or MMSE type operating on a C + 1-restricted space. The equalization performed in reception then certainly requires a matrix inversion, however this matrix is of very limited dimensions compared to the total number N + C reception antennas, especially when N is large. Consequently, the invention advantageously makes it possible to use very simple reception schemes, even when a large number of reception antennas are envisaged. In a particular embodiment of the reception method of the invention, in which C = KR + LR, KR and LR being positive or null integer parameters, the N useful spatial symbols received S '', 1, extracted at the extraction stage check: S '', '= Yn + KR for n = 1,, N.

In a preferred embodiment, the parameters KR, LR and C are chosen so as to ensure, on the one hand, that the focal spot at each receiving antenna (called target) is concentrated for the most part on the target receiving antenna and on C receive antennas around this target receiving antenna (ie negligible signal energy on the other receiving antennas), and secondly, that the focal spot obtained is invariant regardless of the position of the target receiving antenna (ie so as to exploit the "channel hardening" phenomenon described above on a restricted space of dimension C defined around the target receiving antenna (ie total dimension space C + 1 including target receiving antenna).

Thus for this purpose, the parameters C, KR and LR are chosen so that they verify, for at least one so-called target reception antenna, indexed by n and selected from the N + C receiving antennas of the receiver device, a condition focusing the spatial symbol Sn on a space defined by the target receiving antenna indexed by n and KR + LR receiving antennas distributed around the target receiving antenna indexed by n.

According to a first variant of this focusing condition, the parameters C, KR and LR verify that a probability P that: eHnc, in + ef <Ef ', for e integer such that e> LR or -1' <-KR, 1 <-e + n, and 0 5_Ef '<1H, Vn is less than a determined value Pmin, between 0 and 1, where Heq denotes a matrix of dimensions (N + C) x (N + C) resulting of the product of the precoding matrix and of a matrix representing the propagation channel between the transmitting M antennas of the transmitting device and the N + C receiving antennas of the receiving device, Hipq denoting the component situated at the intersection of the jth row and the p-th column of the matrix H. Note that the jth row of the matrix H "corresponds to the received spatial symbol Yi, and the p-th column of the matrix H" corresponds to the spatial symbol S. This ensures that the focal spot has a level of important signal in a C + 1 restricted space around the target antenna (including the target antenna), and a negligible signal level outside this space. According to a second variant of the focusing condition, the parameters C, KR and LR verify that a probability P 'that: fPnoise Zn, n + e 5_Ef', for -e integer such that -e> LR or -e < -KR, 1 <-e + n, and 0 <Efo, <1} is less than a determined Pimin value between 0 and 1, where: 30Pnoise denotes a noise power at the target receiving antenna indexed by not ; - Zn, '+ e denotes the power received at the target receiving antenna indexed by n associated with the spatial symbol S'14. This second variant of the focusing condition makes it possible to ensure that after the step of extracting the N useful spatial symbols from the spatial symbols received, the residual interference between symbols remains negligible in front of the noise. As mentioned previously, the parameters C, KR and LR can also be chosen so as to check a channel hardening condition around the target antenna, aiming to guarantee the invariance of the focal spot regardless of the position of the target. target antenna. This invariance makes it possible to estimate the symbols emitted by the transmitting device, thereby decreasing the number of unknowns to be estimated on reception. To this end, according to an alternative embodiment of this "channel hardening" condition, the parameters C, KR and LR verify that there is a form of focal spot represented by a vector of dimension C + 1.1 V (-LRPv -LR + 1, V KR) of complex numbers, such that, whatever the reception antenna considered of the receiver device indexed by n, the form of a focal spot centered on this reception antenna and represented (- n ## STR2 # n + -e, verifies: d (A ', v) dmax where: dmax is a predetermined value, d (A', y) denotes a distance between the focal spot shapes A11 and y, and H eq denotes a matrix of dimensions (N + C) x (N + C) resulting from the product of the precoding matrix and a matrix representing the propagation channel between the M transmitting antennas of the transmitting device and the N + C receiving antennas of the device récepteu r, Hipq designating the component located at the intersection of the jth row and the p-th column of the matrix H ". It should be noted that the parameters C, KR and LR can advantageously verify both a focusing condition and a channel hardening condition as mentioned above. Since the interference is concentrated on C + 1 receiving antennas, it can then be demonstrated, assuming a perfect estimate of the MIMO propagation channel between the antennas, that the interference between the reception antennas is suppressed after the extraction. N spatial symbols useful in reception. In a particular embodiment, the different steps of the transmission method and the steps of the reception method are determined by computer program instructions. A ', KR) of complex numbers, Accordingly, the invention also relates to a computer program on an information carrier, this program being capable of being implemented in a transmitting device or more generally in a computer, this program comprising instructions adapted to the implementation of the steps of a transmission method as described above.

The invention also relates to a computer program on an information carrier, this program being capable of being implemented in a receiver device or more generally in a computer, this program including instructions adapted to the implementation of the instructions. steps of a reception method as described above. These programs can use any programming language, and be in the form of source codes, object codes, or intermediate codes between source codes and object codes, such as in a partially compiled form, or in any other desirable shape. The invention also relates to a computer-readable information medium, comprising instructions of a computer program as mentioned above.

The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a diskette (floppy disc) or a disk hard. On the other hand, the information medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network. Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question. The invention also relates to a transmitting device comprising a number M of transmitting antennas, able to transmit data symbols to a receiving device comprising a number N + C of receiving antennas, M, N and C being integers greater than or equal to 1, said transmitting device comprising - space frequency multiplexing means of a block of N data symbols D1, D2,..., DN using an inverse discrete Fourier transform, and providing a block of N useful spatial symbols S ', i, Su, 2, .-., Su, N; means for inserting C space redundancy symbols into the block of N useful spatial symbols, providing a block of N + C spatial symbols Si, S2, SN + C means for precoding the block of N + C spatial symbols Si , S2, SN + c provided by the insertion means, using a focusing matrix of dimensions Mx (N + C), providing M precoded spatial symbols X1, X2, ..., Xm; and - means of transmitting in a focused manner each precoded symbol on a transmitting antenna distinct from the transmitting device. Correlatively, the invention also relates to a receiver device comprising a number N + C of receiving antennas, able to receive, on said N + C receiving antennas, N + C spatial symbols Y1, Y2, YN + C; said received spatial symbols resulting from the propagation via a propagation channel of M precoded spatial symbols X1, X2, ..., XM transmitted by a transmitting device via a number M of transmitting antennas, said precoded spatial symbols X1, X2 , ..., XM being derived from a precoding, with the aid of a focusing matrix, of a block of N + C spatial symbols S1, Sz, SN-FC, these spatial symbols being derived from a multiplexing in spatial frequency of data symbols D1, D2, ..., DN using an inverse discrete Fourier transform and an addition of C spatial redundancy symbols, M, N and C being integers greater than or equal to 1, said receiving device comprising - N means for extracting N received useful spatial symbols denoted S '', 1, S ',,, 2, ..., S'ti, N from said N + C received spatial symbols Y1, Y2, ..., YN + C - means for spatial frequency demultiplexing of the N useful spatial symbols received S'u, 1, 5%, 2, ..., S'tJ, N using a trans discrete Fourier ormea, in a block of N demultiplexed data symbols D'2, D'N; and means for equalizing the block of N demultiplexed data symbols D'2, ..., D'N. In a particular embodiment, the N + C receiving antennas form an array of antennas having a geometry such that if we designate by FI, the set of Cartesian coordinates of (C + 1) receiving antennas denoted RXk, RXk +1, .-., RXk + c, there is a translation and / or a rotation allowing to go from F1 to rk whatever k is between 1 and N. For example the N + C receiving antennas can form a network circular or a linear network in which the receiving antennas are evenly spaced.

According to another aspect, the invention relates to a multi-antenna transmission system comprising a transmitter device according to the invention; and a receiver device according to the invention. The transmitting device, the receiving device and the multi-antenna transmission system have the same advantages as those mentioned above for the transmission method and the reception method according to the invention. It may also be envisaged, in other embodiments, that the transmission method, the reception method, the transmitting device, the receiving device and the multi-antenna transmission system according to the invention present in all or part of a combination aforementioned characteristics.

BRIEF DESCRIPTION OF THE DRAWINGS Other features and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate an embodiment having no limiting character. In the figures: FIG. 1 schematically represents a transmission system, a transmitting device and a receiver device according to the invention, in a particular embodiment; FIGS. 2A and 2B show, schematically, the hardware architecture of the transmitting device and the receiving device of FIG. 1; FIGS. 3A and 3B illustrate two examples of antenna network geometries that can be envisaged at the receiver device; FIG. 4 represents the main steps of a transmission method and of a reception method according to the invention, when they are implemented respectively by the transmitting device and by the receiving device of FIG. particular embodiment; FIG. 5 represents an example of parameters C, KR and LR satisfying a focusing condition around a target antenna RX,; and Figure 6 shows an example of parameters C, KR and LR satisfying a channel hardening condition.

DETAILED DESCRIPTION OF THE INVENTION FIG. 1 represents, in its environment, a multi-antenna transmission system 1 of a telecommunications network NW (not shown), according to the invention, in a particular embodiment. The telecommunications network NW is here a mobile telecommunications network, for example offering a mode of communication in a mode TDD (Time Division Duplex). However, this assumption is not limiting, the invention also applies for example to telecommunications networks operating in FDD (Frequency Division Duplex) mode. Moreover, the invention applies both upstream (terminal direction to base station) and downstream (base station to terminal direction) of the wireless telecommunications network NW. According to the invention, the transmission system 1 comprises: a transmitting device 2, according to the invention, comprising M transmitting antennas TX1, TX2, ..., TXM, M being an integer greater than or equal to 1; and a receiver device 3, according to the invention, comprising N + C transmitting antennas RXI., RX2,..., RXN + c, N and C being integers greater than or equal to 1. It is assumed here that N = 2P, where p denotes a positive integer greater than or equal to 1.

The transmitter device 2 and the receiver device 3 are separated by a propagation channel 4. It is assumed here that the channel 4 is a flat frequency channel (flat fading), in other words, all the frequencies are attenuated in a similar way, so that each propagation sub-channel between a transmitting antenna TX ,, and a receiving antenna RX can be modeled by a single complex coefficient (ie a single path). Thus, the channel 4 can be modeled by a matrix of dimensions (N + C) xM of complex components h ',,' with m = 1, ... M and n = 1, ..., N + C, the component hm, denoting the coefficient of the channel between the transmitting antenna TX, and the receiving antenna RX ,. It should be noted that in the case of a frequency selective propagation channel (ie multipath), if a multi-carrier technique is used (for example OFDM), there is a communication system with a flat frequency channel. for each subcarrier, so that the invention applies for each of these systems independently. In the example envisaged here, as described above, one places oneself in a communications context in TDD mode (Time Division Duplex). In such a context, reciprocity between the uplink channel and the downlink channel is generally accepted. In other words, the upstream Channel State Information (CSI) is proving to be the downstream channel and vice versa. Such an estimation can therefore be used in transmission, for example by the transmitting device 2. Alternatively, in a FDD context, a return path between the transmitting device 2 and the receiving device 3 can be implemented so that the transmitting device 2 may have knowledge of the transmission channel, estimated and then raised by the receiving device 3. In the embodiment described here, the transmitting device 2 and the receiving device 3 have the hardware architecture of a computer. Thus, with reference to FIG. 2A, the transmitter device 2 comprises, in particular, a processor 5, a random access memory 6, a read-only memory 7, and communication means 8 enabling it to transmit signals on the communication network NW to destination of other equipment such as in particular to the receiving device 3. These communication means include in particular the M transmit antennas of the transmitting device 2, as well as means for shaping the signals transmitted on the M transmit antennas in accordance with FIG. the communication protocols defined on the NW telecommunications network. The read-only memory 7 of the transmitting device 2 constitutes a recording medium in accordance with the invention, readable by the processor 5 and on which is recorded a computer program according to the invention, comprising instructions for the execution of the steps of an emission process according to the invention.

Similarly, with reference to FIG. 2B, the receiver device 3 comprises, in particular, a processor 9, a random access memory 10, a read-only memory 11, and communication means 12 enabling it to transmit signals on the communication network NW to other equipment, such as for example to the transmitting device 2. These communication means include in particular the N + C receiving antennas of the receiver device 3. The read-only memory 11 of the receiving device 3 constitutes a recording medium according to the invention, readable by the processor 9 and on which is recorded a computer program according to the invention, comprising instructions for carrying out the steps of a reception method according to the invention. In the embodiment described here, the N + C receiving antennas of the receiver device 3 form an antenna array having a geometry such that if the set rk designates the set of Cartesian coordinates f (xk + i, yk + i, zk + i), i = 0, ..., Cl of (C + 1) antennas RXk, RXk + i, ..., RXk + c, there exists a translation and / or a rotation allowing to pass from r1 to rk irrespective of k between 1 and N. Figures 3A and 3B illustrate two examples of geometries verifying such a condition. For the sake of simplicity, these geometries are illustrated in two dimensions.

More precisely, FIG. 3A represents a circular antenna array formed of N + C antennas regularly spaced at the same angle θ, along a circular axis and a direction, the antennas being numbered in this direction in ascending order. In the example shown, N = 16 (i.e. p = 4) and C = 2. For illustrative purposes, three sets F1, r5 and 1 "9 are shown in Figure 3A, these sets respectively form the same geometrical figure with a rotation close to it Figure 3B shows a linear antenna array formed of N + C antennas spaced The antennas are numbered according to this direction in ascending order, in the example shown, N = 16 (ie p = 4) and C = 2. For illustrative purposes, two sets r 1 and r 1, 3B These assemblies respectively form the same geometrical figure with a translation close.We will now describe, with reference to FIG. 4, the main steps of a transmitting method and a reception method according to FIG. according to the invention, in a particular embodiment in which they are respectively implemented by the transmitting device 2 and the receiving device 3 of FIG. transmission described here is a single-carrier scheme. In the embodiment described herein, for the sake of simplification, only the processing of a symbol is used to illustrate the steps of the transmission method and the reception method. However, the invention obviously makes it possible to transmit symbol flows, the steps described being then performed for each symbol of the flow.

We will firstly describe the main steps of the transmission method according to the invention implemented by the transmitter device 2 (steps grouped under the general TX PROC step).

According to the invention, a block of N data symbols, denoted D = (D1, D2,..., DN), is first multiplexed into spatial frequency (step E10), ie each Dn symbol is associated with a spatial frequency fs (n), The data symbols D 'n = 1, ..., N are here symbols with complex values of any constellation (eg QPSK (Quadrature Phase Shift Keying) or QAM (Quadrature Amplitude) Modulation)). The invention is of course also applicable to real-valued symbols, for example derived from a BPSK (Binary Phase Shift Keying) modulation. To illustrate the notion of spatial frequency fs, let us consider a real and continuous signal SiG (x) received at a position x, where x denotes a curvilinear coordinate expressed in meters. The signal SiG (x) is a spatial signal, which can be decomposed into spatial frequencies fs in the following form: 1 + FS12 SiG (x) A (fs) cos (2n- fs x 1) (fs) ) dfs Fe co FS LE.512 where A (fs) and si) (fs) are coefficients resulting from the decomposition of the signal SiG (x) into spatial frequencies. We therefore note that the notion of "spatial frequency" is a dual relation of the notion of space; it does not have a relationship strictly speaking with any notion of time or frequency. To illustrate now the concept of spatial frequency in the discrete and complex domain, Ri denotes the discrete spatial sample received at the reception antenna n positioned spatially at the coordinate xn. It is assumed that there are N distinct receiving antennas and thus N distinct discrete spatial samples. Any sample Rn can be decomposed into N spatial frequencies using an inverse discrete Fourier transform, as follows: p = N 1 .2Tr (n-1) (p-1) R = Upe rr VN p = 1 each spatial frequency carrying a complex coefficient Up, p = 1, ..., N. In other words, the principle of spatial frequency multiplexing consists in associating the symbols to be transmitted Dn, n = 1,..., N on N spatial frequencies, that is to say to choose: Up = Dp, p = 1, ..., N. It should be noted that this principle of multiplexing in spatial frequency is different from the principle of prior art multiplexing known from the state of the art, which consists in directly associating the symbols to be transmitted on N reception antennas and therefore in not using frequency decomposition. spatial (ie Rp = Dp, for p = 1, ..., N). Spatial frequency block D multiplexing is performed from an inverse Fourier transform, and more specifically in the embodiment described herein using a N-dimensional inverse inverse (IDF) or inverse (IDFT) transform. = 2P. Thus, if AIDFT denotes the matrix of dimensions NxN of this inverse Fourier transform, each component Aili ° 7 of this matrix, n, m = 1, ..., N is defined by: AIDFT - 1 e + / 27, (n -12Pm -i) ne .N / 2P In other words, during the spatial frequency multiplexing step E10, the transmitting device 2 calculates a block of N so-called useful spatial symbols denoted Su such that Su A ID FT D Thus, the spatial frequency multiplexing step E10 involves associating each data symbol Dn, n = 1, ..., N at a spatial frequency, using an IDFT operation. The symbols resulting from this operation are spatial symbols, each spatial symbol being associated with a receiving antenna. As mentioned above, this step differs from the state of the art, in which each data symbol Dn is directly associated (i.e. without decomposition) with a reception antenna. Then, according to the invention, the transmitting device 2 adds to the block Su of useful spatial symbols, C space redundancy symbols (step E20), in the form of a prefix of KT space symbols of redundancy, added at the head of the block Su, and a suffix of LT spatial symbols of redundancy, added following the block S ', with LT + KT = C. It will be noted that LT and KT denote any two integer parameters greater than or equal to 0 such that LT + KT = C. For example, we can take KT = C, LT = 0 (adding a prefix only), or KT = 0 and LT = C (adding a suffix only). In the example envisioned here, the prefix and suffix added to the block of useful spatial symbols Su are cyclic, ie the added prefix is made up of the last KT useful space symbols of Su, and the added suffix is made up of the first useful space symbols LT. from Su. Thus, if Z denotes the given N + C block obtained at the end of this redundancy step, we have: SKT + p = S ', p, for 1 5_p 5 N, for the central part. Sp = Su, N-KT + p for 1 5_ p 5_ KT, for the prefix SKT + N + p = Su, p for 1 5_ p <LT, for the suffix. where Sp, respectively Su, p, denotes the p-th symbol of the block of spatial symbols 5, respectively of the block of useful spatial symbols S '. The block S of N + C spatial symbols is then precoded, using a focusing precoding matrix Q (focusing matrix in the sense of the invention), of dimensions Mx (N + C) (step E30). . This precoding step makes it possible to focus each spatial symbol S, towards a reception antenna RX, of the receiver device 3, and results in a block X of M precoded spatial symbols such that: X = QS The focusing matrix Q is, in the mode embodiment described here, a time reversal matrix, defined, in a manner known per se, by: Q where H denotes the Hermitian operator and 11 denotes the matrix of the estimated coefficients of the MIMO propagation channel 4, between the M antennas; transmission of the transmitter device 2 and the (N + C) receiving antennas of the receiver device 3. The matrix Ili is obtained using propagation channel estimation techniques 4 known to those skilled in the art and not described here for example by issuing pilot symbols. In the example considered here, one places oneself in a TDD context where the reciprocity of the propagation channel between the uplink and the downstream channel is generally accepted, as previously described. The estimate of the channel available at the transmitter device 2 is therefore derived from an estimation made by the transmitter device 2 during a communication implemented in the device 3 to device 2 direction.

In a variant, when the channel is not reciprocal between the uplink and the downlink (for example, in an FDD context), the estimate of the channel 4 can be carried out by the receiver device 3 and sent via of a return path, to the transmitting device 2. The block X = (Xi, X2,..., Xm) of M precoded spatial symbols is then transmitted on the propagation channel 4 via the M transmit antennas TX1. TX2, ..., TXm of the transmitter device 2.

More precisely, each precoded spatial symbol Xll 'of the block X is transmitted on the antenna TXm, m = 1, ..., M. The symbols emitted by the TXm antennas propagate in the propagation channel 4 and are received on the N + C receiving antennas RXi, RX2, RXN, c of the receiver device 3. We will first describe the main steps of the method receiver according to the invention implemented by the receiver device 3 (steps grouped under the general step RX PROC). The receiving device receives on its N + C receiving antennas the signal Y = (Y1, Y2, ..., YN + c) composed of (N + C) space symbols received, and distributed on the N + C receiving antennas. The block of (N + C) spatial symbols received Y is such that: Y = HQ S + No Heq where: - Nel denotes the matrix of dimensions (N + C) x (N + C) of the equivalent channel between the block of M emitted spatial symbols Set the (N + C) spatial symbols block received Y, after precoding by the focusing matrix Q and propagation in the propagation channel 4, and - No denotes a vector of dimension N + C, representing a white noise Gaussian additive at the receiving device 3. According to the invention, the receiver device 3 extracts N useful spatial symbols received from the N + C spatial symbols received from the signal Y (step E40), to form a block S'u =, S ',, N) of N received useful spatial symbols. More particularly, during this extraction step, it carries out the following operation: S u, n = YKR + n for 1 <n <N, KR being a positive or zero integer parameter such that KR + LR = C, LR is also a positive integer parameter or zero. This step consists of absorbing inter-antenna interference.

In the embodiment described here, the parameters C, KR and LR are advantageously chosen so as to verify two conditions, namely: a first focusing condition, around a receiving antenna RX, target (and more precisely on C = KR + LR reception antennas distributed around the target antenna), selected from the reception antennas RXi, RX2, -, PXN, c of the receiver device 3, of the spatial symbol Sr, transmitted by the transmitter device 2 to the target antenna RXn; and a second so-called channel hardening condition, to ensure that the focal spot formed by this symbol around this target receiving antenna is invariant or substantially invariant regardless of the position of the target antenna. It is recalled that a focal spot characterizes the level of the signal intended for this reception antenna, received on the receiving antennas of the receiving device. In the embodiment described here, the first focusing condition consists in determining the parameters C, KR and LR so that a probability P that: Hielqn + e Heq n, n is less than a determined value Pmin, between 0 and 1. The index n denotes the index of the target receiving antenna RXn considered. Ef 'is preferably small, for example Ef oc = 0.1. This condition makes it possible to ensure that outside of a restricted space defined by the target receiving antenna and C antennas distributed around the target antenna (KR antennas before, and LR antennas after the target antenna), the power of the received signal is small compared to the power of the signal received at the target antenna RX, - ,.

Alternatively, this first focusing condition can be formulated differently. For example, it is possible to determine the parameters C, KR and LR so that a probability P 'that: Ef', for e integer such that e> LR or e <-KR, 1 <n + -e and 0 <Efoc <1 fZ n, n + -e Pnoise 5_ E foc, for e integer such that e> LR or e <-KR, 1 <e + n, and 0 <Efoc <1} is less than a value Pmin determined, between 0 and 1, where: - n denotes the index of the target receiving antenna considered RXn; P'm 'designates the noise power at the target receiving antenna RXn; Z77, n + e denotes the power received at the target receiving antenna RX 'associated with the spatial symbol S' + e. Ef 'is preferably small, for example Ef' = 0.1. The signal-to-noise ratio can be estimated in a manner known per se, using pilot symbols. FIG. 5 illustrates an exemplary TFoc (n) focal spot obtained on a target antenna RXn of the receiver device 3, when the first focusing condition is verified. This focal spot is centered on the target receiving antenna RX, and is distributed on KR = LR = 2 antennas on both sides of the target antenna. In this figure Lo designates the signal power level below which the signal is considered negligible in relation to the noise of the receiver device. In the embodiment described here, the second channel hardening focusing condition is to determine the parameters C, KR and LR so that there is a focal spot shape represented by a dimension vector C + 1 y = (y_ LR, -LR + 1, vi (R) complex numbers such that, regardless of the target receiving antenna RX, considered the receiving device, the shape of a focal spot centered on this receiving antenna and --n, -LR, An, -LR + 1) ---, A ', KR) of numbers represented by a vector of dimension C + 1 An = (A complexes with A', '+ e = leign_ He for -L <-e <K, check: d (An, v) 5 dmax where: - dmax is a predetermined value, and independent of n - d (A ', y) denotes a distance between the focal spot forms A17 This distance can be for example the ratio between the quadratic norm of the vector (A77-y) on the quadratic norm of the vector U. The distance dmax is then between 0 and 1 (for example dmax = 0.1). It should be noted that the determination of the parameters C, KR and LR can be carried out as a step prior to the steps implemented in the reception method, for example by means of simulations.

Figure 6 illustrates the channel hardening condition on C antennas around the target antenna and the invariance of the focal spots as a function of the position of the target antenna. According to the invention, the block of useful spatial symbols received S 'extracted in step E40 is then demultiplexed during a step of demultiplexing in spatial frequency (step E50), performing the reverse operation of the step E10. This step is carried out using a Discrete Fourier Transformed P110 or DFT of dimension N = 2P. Thus, if ADFT denotes the matrix of dimensions NxN of this Fourier transform, each component AiD, FinT of this matrix, n, m = 1, ..., N is defined by: 1 j2Tr (n-1) (n- 1) e ADFT 2P 2P In other words, during the step of spatial frequency demultiplexing E50, the receiver device 3 calculates a block of N data symbols denoted D '= (D'1, D'2,, D' N) Such as: ADFTst u The block D 'of N data symbols is then equalized (step E60). For this purpose, the receiver device 3 considers that, following the steps E10 to E50 previously described, the block of data symbols D 'can be written in the following form: ADFT AIDFT D ADFT m circ v 0, u deq where: liceigrc denotes the matrix of the "folded" equivalent channel of size NxN, in the spatial frequency domain (i.e., the equivalent channel matrix between the useful space symbol block Su and the useful space symbol block received SU ). For n = 1, ... N, and m = 1, ... N, each component Hcirc, r1,7n of the matrix He9ctrc is given by: H ceiqrc, n, in Hue 1,77, e (H, rn, m) with g (I-itti, m) = 0 if N + m> N + C, and g (Hue% m) = Hul, m + N if N + m5..N + C. f I, "is the size matrix Nx (N + C) of the equivalent channel between the block of spatial symbols S and the block of useful spatial symbols received S11 defined by the following coefficients: ueq eq" u, n, p "uKR + n, p for 1 n5_Net 1 <p <N + C - No, u denotes the noise extracted from the signal received during step E40 with the following coefficients: NO, u, n = No, KR + 'for 1 <n <N. In practice, this assumption is verified especially when the parameters C, KR and LR satisfy the above mentioned focusing and channel hardening conditions, otherwise it is an approximation made by the receiver device 3 in order to simply equalize the signal D 'It is noted that, under the assumption of a perfect estimate of the channel, the matrix .deq is diagonal.

This matrix can in practice be estimated using pilot signals in a manner known per se. Thus, for example, such a method may consist in processing a Dpiiote block of N pilot symbols known from the transmitter device 2 and the receiver device 3, according to the steps E10 to E50 presented above. Each component A7e, q, n = 1, ..., N of the diagonal of the matrix Jeq is then obtained in a very simple way by:, aeq .D'pilot, n Dpilot, n if Driver designates the block of useful pilot symbols received at the end of step E50. The receiver device 3 then estimates the block D of the transmitted data symbols by equalizing D 'by the diagonal matrix Den, without resorting to a matrix inversion, but by conventional methods known to those skilled in the art, exploiting the fact that Den is diagonal. For example, each data symbol b ', n = 1, ..., N of block b is computed by:, D' D, = 1 11 It is noted that if the focusing conditions (eg the first condition of focussing) and channel hardening mentioned previously are verified, for example for £ f oc Ehard = 0 and Pim = 1 to simplify calculations, we can show that the matrix Hceigrc exists and verifies: diagonalisableecid Without a Fourier basis, that is that is to say that there exists a diagonal matrix Aeg such that: ueq = AIDFT AeqADFT "circ" On the other hand: AIDFT ADFT "=" I, where I designates the identity matrix Aeg therefore corresponds to the matrix of the equivalent channel in the domain of spatial frequencies Furthermore: = ADFT sfu, ._. Aeq D + ADETN0, u The equalization step which consists of calculating D '' D11 = a 11 thus leads, assuming that the noise is zero in b '= D. In other words, the interference between symbols is perfectly eliminated.The invention applies generally to comm wireless unications.

In the embodiment described here, the signal transmitted on each transmitting antenna is a single-carrier signal with a flat frequency channel between each transmitting and receiving antenna. The invention can however also be applied for the transmission of a multi-carrier signal: in this case, the different steps of the transmission method and of the reception method are implemented for each sub-carrier of the multi-carrier signal . Thus, the invention can be advantageously applied in a plural manner to a plurality of narrow frequency bands; it can, for example, apply to the OFDM: indeed, a sub-carrier T OFDM system can be seen as a set of independent T systems whose channel is flat frequency. For each of these systems, the described solution can be applied. The invention also applies to OFDM-based technologies, especially as defined in the LTE (Long Term Evolution) standard.

Claims (17)

REVENDICATIONS1. A method of transmitting data symbols by a transmitting device (2) having a number M of transmitting antennas, to a receiving device (3) having a number N + C of receiving antennas, M, Wherein N and C are integer parameters greater than or equal to 1, said transmitting method comprising: a step (E10) of spatial frequency multiplexing of a block of N data symbols DI, D2, ..., DN resulting in a block of N useful spatial symbols 5u, I, Su, 2, -, Su, N, this step using an inverse discrete Fourier transform (IDFT); a step (E20) of adding C spatial redundancy symbols to the block of N useful spatial symbols, resulting in a block of N + C spatial symbols S1, Sz, SN + C; and a step (E30) of precoding the N + C spatial symbol block S1, Sz, SN-rc by means of a focusing matrix of dimensions Mx (N + C), said precoding step providing M symbols the precoded spatial symbols X1, X2, -, XM, each precoded spatial symbol being transmitted on a transmitting antenna distinct from the transmitting device. 2. A transmission method according to claim 1 wherein C = KT + LT, LT and KT being positive or null integer parameters, the step (E20) of adding C spatial redundancy symbols comprising: - the insertion a suffix of LT symbols following the block 5 ', 1.5u, 2, ..., Su, N of N spatial frequency multiplexed data symbols; and / or - the insertion of a prefix of KT symbols at the head of the block Sio, Su, 2, ..., Sum of N spatial frequency multiplexed data symbols. The emission method according to claim 2, wherein: KT = C and LT = O; or KT = O and LT = C. 4. Transmission method according to claim 1, wherein the focusing matrix is a time reversal matrix obtained from an estimate of a propagation channel matrix (4) between the M transmit antennas of the device. transmitter and the N + C receiving antennas of the receiver device. A computer program comprising instructions for performing the steps of the transmitting method according to claim 1 when said program is executed by a computer. 6. Reception method, by a receiver device (3) comprising an N + C number of reception antennas, N + C spatial symbols Y1, - Y 2, - - -t YN + C, said received spatial symbols resulting the propagation via a propagation channel (4) of M precoded spatial symbols X1, X2,..., Xm transmitted by a transmitting device (2) via a number M of transmitting antennas, said precoded spatial symbols X1, X2, ..., Xm being derived from a precoding, using a focusing matrix, of a block of N + C spatial symbols S1, Sz, SNA-C these spatial symbols SI, S2, - -t SN-Fc resulting from a spatial frequency multiplexing of data symbols D1, D2,..., Drsi using an inverse discrete Fourier transform, and an addition of C spatial redundancy symbols, M, N and C being integer parameters greater than or equal to 1, said receiving method comprising: - a step (E40) of extracting N useful spatial symbols received S'u, i, - .., S'u, N p armi said N + C received spatial symbols Y1, Y2, YN + C; a spatial frequency demultiplexing step (E50) of the N received spatial useful symbols S '', 1, Siu, N using a discrete Fourier transform, resulting in a block of N demultiplexed data symbols D'1, D'N; and a step of equalizing the block of N demultiplexed data symbols D'1, D'2, D'N. 7. The reception method according to claim 6, wherein C = KR + LR, KR and LR being positive integer parameters or zero, and in which the N useful spatial symbols received S '', 2, extracted in step d extraction verify: S ',' = Yn ÷ KR for n = 1, ..., N. 8. The reception method as claimed in claim 7, wherein the parameters C, KR and LR satisfy, for at least one so-called target reception antenna, indexed by n and selected from the N + C receiving antennas of the receiver device (3), a focus condition of the spatial symbol 5 ,, on a space defined by the target receiving antenna indexed by n and KR + LR receiving antennas distributed around the target receiving antenna indexed by n. 9. The reception method according to claim 8, wherein the parameters C, KR and LR verify that a probability P that: f 11:, qn + e Ef ', for? integer such as? > LR or? <-KR, 1 <-e + n, and 0 Ef '<1} eq is less than a determined value Pmin, between 0 and 1, where H'q denotes a matrix of dimensions (N + C) x (N + C) resulting from the product of the focusing matrix and a matrix representing the propagation channel between the transmitting antenna M of the transmitting device and the N + C reception antenna of the receiving device, H; pq denoting the component located at the intersection of the jth line and the p-th column of the matrix H. The reception method according to claim 8, wherein the parameters C, KR and LR verify that a probability P 'that: Ef', for e integer such that e> LR or -e <-KR, 1 <-e + n, and 0 <Ef '<11Zn, n + e Pnoise is less than a determined value P'min, between 0 and 1, where: Pnoise denotes a noise power at the target receiving antenna indexed by not; Z ',, i + e denotes the power received at the target receiving antenna indexed by n associated with the spatial symbol Sn.i.f. 11. The reception method according to claim 7, wherein the parameters C, KR and LR verify that there is a form of focal spot represented by a vector of dimension C + 1 (Y-LR, V-LR-F). VKR) of complex numbers such that, whatever the reception antenna considered of the receiver device indexed by n, the form of a focal spot centered on this reception antenna and represented by a vector of dimension C + 1 A '= --- An, KR) of complex numbers with A', n ÷ e 117eLein + i, for -LR <-e <KR, 1 <n +, satisfies: (Mn, V) dinax where: 20 d , '' is a predetermined value, d (A, 'y) denotes a distance between the focal spot forms A' and y; and Heq denotes a matrix of dimensions (N + C) x (N + C) resulting from the product of the precoding matrix and a matrix representing the propagation channel between the M transmitting antennas of the transmitting device and the N + C receive antennas of the receiver device, Hypq denoting the component located at the intersection of the jth row and the p-th column of the matrix Heq A computer program comprising instructions for performing the steps of the receiving method according to claim 6 when said program is executed by a computer. 30 Transmitting device (2) comprising a number M of transmitting antennas, able to transmit signals formed of data symbols to a receiving device (3) comprising an N + C number of reception antennas, M, N and C being integer parameters greater than or equal to 1, said transmitting device comprising: space frequency multiplexing means of a block of N data symbols DI, D2,..., DN, said means using an inverse discrete Fourier transform and providing a block of N useful spatial symbols 5 ,,, 1, S ,,, 2, ..., Su, N; means for adding C space redundancy symbols to the block of N useful spatial symbols, providing a block of N + C spatial symbols 51, Sz, SN + C; - Precoding means of the block of N + C useful spatial symbols Si, S2, SN + c provided by the insertion means, using a focusing matrix of dimensions Mx (N + C), providing M precoded spatial symbols X1, X2, ..., XM; and means for transmitting each precoded spatial symbol on a transmitting antenna distinct from the transmitting device. 14. Receiving device (3) comprising a number N + C of receiving antennas, capable of receiving, on said N + C receiving antennas, N + C spatial symbols YY-Y2, -YN + C r, said spatial symbols received due to the propagation via a propagation channel (4) of M precoded spatial symbols X1, X2, ..., XM transmitted by a transmitting device (2) via a number M of transmitting antennas, said precoded spatial symbols X1, X2, ..., XM being derived from a precoding, using a focusing matrix, of a block of N + C spatial symbols SI, S2, ..., st ,,, c , these spatial symbols S1, S2,..., SN ÷ c being derived from a spatial frequency multiplexing of data symbols D1, D2,..., DN using an inverse discrete Fourier transform, and an addition of spatial redundancy symbols, M, N and C being integer parameters greater than or equal to 1, said receiver device comprising: means for extracting N useful spatial symbols received ss - .., S'u, N among said N + C received spatial symbols Y1, lt Y 2r ..., YN + C - spatial frequency demultiplexing means of the N received useful spatial symbols S ',,, 2 , Siu, N, using a discrete Fourier transform and providing a block of N demultiplexed data symbols D'2, D'N; and means for equalizing the block of N demultiplexed data symbols D'1, D'2, D'N. 15. receiving device according to claim 14 wherein the N + C receiving antennas form an antenna array having a geometry such that if we denote by rk the set of Cartesian coordinates of (C + 1) receiving antennas RXk noted , RXk, c, there is a translation and / or a rotation allowing to pass from ri to rk whatever k is between 1 and N. The receiver device of claim 15 wherein the N + C receive antennas form a circular array or a linear array and are evenly spaced apart. 17. Mufti-antennas transmission system comprising: a transmitter device according to claim 13; and a receiver device according to claim 14.